WHITE PAPER Guide to Measuring Water in Oil We have all heard the saying, “Oil and water don’t mix.” Unfortunately that doesn’t necessarily apply to lubrication oils. Water can exist in several states in lubrication oils and can do quite a bit of damage to valuable assets if left unchecked. In this guide we explore the challenges posed by water in lubrication oils and discuss the methods available to reliability professionals for measuring water in oil. Background Water contamination in industrial oils can cause severe issues with machinery components. The presence of water can alter the viscosity of a lubricant as well as cause chemical changes resulting in additive depletion and the formation of acids, sludge, and varnish. Water testing is always a part of any lubricant condition monitoring program. Water contamination in industrial oils with strong water separation properties has been historically difficult to measure with any technique. Water can coexist as three phases in an oil sample: dissolved, emulsion mix, and free globules which makes it challenging to obtain a representative sample. The additive formulations in these types of oils are designed to separate water efficiently above saturation limits of 50-250 ppm water. Above saturation, the water will form discrete water droplets and eventually separate from the oil if left to sit (Figure 1). Dissolved water is typically not Figure 1. Sample of used Chevron GST 32, too much of a concern for industrial applications but free as received by the lab after shipment water or emulsified water must be stringently controlled. from a power generation plant. What techniques are available? ■ INFRARED A most promising technique for measuring water contamination is infrared spectroscopy. It is a widely used and accepted chemical-free measurement. In a very general sense, FluidScan spectroscopy is the study between the interaction of radiated energy and matter. A spectrometer consists of a radiative source, a detector, and a computer or other converter of the detector signal to useful information. The sample to be studied is placed between the radiative source and the detector. In infrared absorption spectroscopy, the incident light beam is passed through a sample and the transmitted light is collected by the detector and typically reported as a spectrum of the transmitted or absorbed light as a Figure 2. Schematic of typical spectrometer function of the wavelength, , of the incident beam. Guide to Measuring Water in Oil | 2 Pure water absorbs infrared light and can be detected by a peak in the infrared spectrum at around 3400 cm-1. When water is dissolved in another medium like oil, a water molecule still absorbs infrared light although the peak might be shifted slightly due to the different environment around that water molecule. ASTM standard practice E2412 measures water by correlating the peak in the infrared spectrum around 3350 cm-1 to the concentration of water (Figure 3). This method works great for certain types of applications, such as for motor oils which contain additives to help dissolve water. The analysis is fast, easy, and very reliable; results can be achieved by even a non-technical user. When water exists as free or emulsified in oil, it can cause the light passing through Figure 3. Measurement for water in ASTM E2412. the oil sample to scatter, thus affecting the measurement. In industrial lubricants, most of the water exists as an emulsion or as free water due to the demulsibility additives present in the oil. ADVANTAGES: Water droplets suspended in oil can scatter • Correlates well to Karl Fischer light instead of absorb it. • Portable and easy to Thus a water absorption peak around use -1 • Requires only one drop 3400 cm cannot easily be correlated of oil to the concentration of water above the • Results in less than 2 saturation limit. One successful approach minutes to counteract this scattering is the water • Can measure other oil parameters, such as stabilization method.1 This method pretreats TAN, TBN, oxidation, the sample with an additive, such as a sulfation surfactant, which aids in the dissolution of DISADVANTAGES: • More expensive the water in the oil so that the traditional than some other test water peak can be detected by infrared methods spectroscopy. The water stabilization method can work well with a skilled operator, but requires precise addition of chemicals into an oil sample prior to analysis- and this is not readily available in the field when needed. For this reason, the method has not been received well by condition monitoring labs using FTIR as a screening tool or by the end user in the field. Another approach is the use of a homogenizer to mechanically Figure 4. Graphical representation of light scattering in used turbine oil due to varied water droplets. Spectrum A is a used turbine oil incorporate the water as discrete water droplets suspended in with 29,000 ppm water contamination immediately analyzed after the oil. Samples homogenized with a CAT 120X homogenizer homogenization. Spectrum B is a used turbine oil with 9,500 ppm water and allowed to sit at room temperature for at least a minute contamination immediately analyzed after homogenization. Spectrum demonstrated excellent correlation to Karl Fischer coulometric C is the same sample as in A (29,000 ppm) but has been allowed to sit for 45 minutes after homogenization. The change in concentration and titration. The degree of elastic light scattering caused by a water droplet size is apparent in the degree of baseline lift. water-in-oil mixture indeed depends on the concentration of water present, but it also is strongly influenced by how the water is physically dispersed in the oil: the number and size of discrete water droplets present in the oil (Figure 4). The homogenizer acts to create a repeatable distribution of water droplets uniformly dispersed in the oil. Then, the degree of light scattering can be measured reliably by an infrared spectrometer. The FluidScan oil analyzer is a rugged, handheld IR analyzer used to measure oil condition and chemistry. A series of oil samples with varying water concentrations were prepared using the homogenization technique and measured on the FluidScan. The results were compared to those obtained using Karl Fischer coulometric titration as showing in Figure 5. The FluidScan readings show excellent correlation to the titration method. Figure 5. Correlation of FluidScan IR readings The new FluidScan method for analysis of water contamination in turbine oils is a robust, to Karl Fischer reliable method capable of providing immediate alert of severe water contamination. The Guide to Measuring Water in Oil | 3 largest contributor to the variation is the sampling. The homogenizer is a key component of the method. Hand-shaking is not sufficient for obtaining a homogeneous sample and reliable results for water measurement on the FluidScan. Immediate analysis at-site or the preparation of samples prior to analysis with a commercially available homogenizer is recommended for the best results. With best practice sampling techniques, results correlating within 20% to Karl Fischer can be achieved. ■ CRACKLE TEST CRACKLE TEST ADVANTAGES: The Crackle Test is a simple test to determine whether or not there • Inexpensive is water in lubrication oil. The Crackle Test can be performed in • Quick and easy • Can be done on-site No visible or audible change: the field using a hot plate. The Crackle Test is mostly qualitative. No free or emulsified water It tells you whether there is water present or not but doesn’t DISADVANTAGES: Not quantitative measure the amount of water present. Careful observation of the • • Safety concerns oil on the hot plate can give some semi-quantitative information Very Small Bubbles (0.5 mm) produced • Inconsistent results and quickly disappear: about the amount of water present, but it is very dependent on the between operators 0.05 - 0.1% 500-1000 ppm skill of the operator. The Crackle Test requires a temperature controlled hot plate. Typically the hot plate Bubbles approximately 2 mm are produced, temperature is set to approximately 320°F (160°C). Ideally the sample is mixed by gather to center, enlarge to 4mm, vigorous shaking. This creates a more homogeneous suspension of water in oil. Once the and disappear quickly: 0.1 - 0.2% 1000-2000 ppm sample is prepared, a drop of oil is placed onto the hot plate and observed. If there is no water in the oil, no “crackling” will be observed. Conversely if the oil does contain liquid water, the heat will cause it to change to the vapor stage creating Bubbles 2-3 mm are produced growing to 4 mm, process repeats, possible violent observable bubbles in the oil droplet. The size of the bubbles roughly corresponds to bubbling and audible crackling: the amount of water present in the oil; the larger the bubbles, the more water there is 0.2 and more >2000 ppm dissolved in the oil. By observing the size of the bubbles created on the hot plate, some Fitch, J. (1998). Oil Analysis for Maintenance Professionals, Tulsa OK, Noria Corp. quantitative judgements can be made about the concentration of water present in the oil sample. ■ CALCIUM HYDRIDE TEST KITS ADVANTAGES: Another way to determine water concentrations in the field is • Portable by using a calcium hydride test kit. In this test method, a known • Relatively easy volume of oil is placed in a sealed container with a known amount • Inexpensive • Provides quantitative of calcium hydride. The container is shaken vigorously causing results the water in the oil to react with the calcium hydride, producing DISADVANTAGES: hydrogen gas. This chemical reaction is as follows: • Requires chemicals and solvents CaH + 2 H O → Ca(OH) + 2 H 2 2 2 2 • Safety – vessel under The calcium hydroxide [Ca(OH) ] product is a solid and precipitates pressure 2 • Several runs may be out, while the H2 is hydrogen gas and its pressure is measured required depending on using a manometer.